The current successive approximation analog-to-digital converter (ADC) converts an analog signal into an N-bit digital signal using a binary search. The successive approximation ADC includes a comparator. In order to convert an analog signal into an N-bit digital signal, the comparator will execute N times of comparison, and each time of comparison will generate one bit of the digital signal. After N times of comparison are completed, the N bits of the digital signal will be generated sequentially from the most significant bit to the least significant bit. However, when the comparison of one particular bit is very close, a longer time will be required to interpret the comparison result, and then a metastability state may occur. Since the successive approximation ADC in the metastability state may have inaccurate interpretation of the bit or may not have sufficient time to interpret the following bits, the N-bit digital signal converted by the successive approximation ADC may have remarkable error. Therefore, it has become a prominent task for the industries to provide a method for determining the validity of the N-bit digital signal and avoiding the occurrence of metastability state which may cause errors.